CODEWITHSUNDEEP
rplotggplot2heatmap
nik
nik

Reputation: 2584

how to make different color within a heatmap

Here is my data 

df<- structure(list(name = structure(c(2L, 12L, 1L, 16L, 14L, 10L, 
9L, 5L, 15L, 4L, 8L, 13L, 7L, 6L, 3L, 11L), .Label = c("All", 
"Bab", "boro", "bra", "charli", "delta", "few", "hora", "Howe", 
"ist", "kind", "Kiss", "myr", "No", "TT", "where"), class = "factor"), 
    value = c(1.251, -1.018, -1.074, -1.137, 1.018, 1.293, 1.022, 
    -1.008, 1.022, 1.252, -1.005, 1.694, -1.068, 1.396, 1.646, 
    1.016)), .Names = c("name", "value"), class = "data.frame", row.names = c(NA, 
-16L))

I want to have very clear shade separation. if I want to make a heatmap with specific color, I do like this 

ggplot(df, aes(x = 1,y = name, fill = value)) + 
  geom_tile() + 
  ylab("") +
  scale_fill_gradient(low='red', high='blue')

It makes it red to blue. what I want is to make the first 7 rows for example red to blue and the rest pink to black.

Is this possible ?

Upvotes: 0

Views: 1736

Answers (1)

jdobres
jdobres

Reputation: 11957

ggplot generally encourages one scale per data series, but if you really wanted different scales for different subsets of the values, here's one way to do it.

You say you want a different scale for the first 7 rows. Again, tying presentation to the data's position isn't a great practice, but let's suppose we create a "group" column. It's important that this column be numerical, since we'll be using it to "cheat" the values in the heatmap. So a set of group values that separate your distribution would be good:

library(dplyr)
df <- arrange(df, name) %>% 
    mutate(group = c(rep(-3, 7), rep(3, 9)))

The fill aesthetic then becomes value - group, and we use scale_fill_gradientn to assign colors:

ggplot(df, aes(x = 1,y = name, fill = value - group)) + 
    geom_tile() + 
    ylab("") +
    scale_fill_gradientn(colours = c('black', 'pink', 'red', 'blue'))

Which gives you this plot: enter image description here

Upvotes: 1

Related Questions